Edit model card

cls-comment-phobert-base-v2-v2.4.1

This model is a fine-tuned version of vinai/phobert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3274
  • Accuracy: 0.9268
  • F1 Score: 0.8919
  • Recall: 0.8944
  • Precision: 0.8897

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 64
  • eval_batch_size: 64
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • training_steps: 4000

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Score Recall Precision
1.7049 0.96 100 1.4950 0.4614 0.1080 0.1681 0.2101
1.3066 1.91 200 1.0451 0.6598 0.2493 0.2970 0.2150
0.9457 2.87 300 0.7491 0.7972 0.5219 0.5230 0.5238
0.6975 3.83 400 0.5574 0.8497 0.5700 0.5935 0.7143
0.5187 4.78 500 0.4681 0.8665 0.6685 0.6592 0.7077
0.4183 5.74 600 0.4121 0.8821 0.7747 0.7478 0.8761
0.3323 6.7 700 0.3488 0.9040 0.8505 0.8391 0.8647
0.2705 7.66 800 0.3179 0.9124 0.8680 0.8694 0.8683
0.229 8.61 900 0.3109 0.9160 0.8739 0.8778 0.8704
0.1964 9.57 1000 0.3028 0.9175 0.8776 0.8813 0.8741
0.1771 10.53 1100 0.3032 0.9181 0.8807 0.8877 0.8743
0.1518 11.48 1200 0.3151 0.9166 0.8762 0.8702 0.8828
0.1368 12.44 1300 0.2938 0.9214 0.8794 0.8800 0.8789
0.1116 13.4 1400 0.2971 0.9205 0.8795 0.8815 0.8776
0.1136 14.35 1500 0.3011 0.9235 0.8858 0.8825 0.8894
0.094 15.31 1600 0.2937 0.9268 0.8891 0.8933 0.8855
0.0905 16.27 1700 0.3049 0.9265 0.8850 0.8819 0.8886
0.0838 17.22 1800 0.3061 0.9244 0.8823 0.8869 0.8784
0.0749 18.18 1900 0.3275 0.9205 0.8771 0.8839 0.8717
0.0686 19.14 2000 0.3092 0.9295 0.8915 0.8990 0.8846
0.0669 20.1 2100 0.3168 0.9250 0.8836 0.8849 0.8825
0.0582 21.05 2200 0.3339 0.9235 0.8763 0.8926 0.8631
0.0516 22.01 2300 0.3274 0.9268 0.8919 0.8944 0.8897
0.0543 22.97 2400 0.3230 0.9295 0.8913 0.8882 0.8946
0.0435 23.92 2500 0.3364 0.9253 0.8806 0.8705 0.8918
0.0405 24.88 2600 0.3492 0.9241 0.8816 0.8821 0.8819
0.0398 25.84 2700 0.3558 0.9238 0.8799 0.8796 0.8807
0.0363 26.79 2800 0.3605 0.9223 0.8742 0.8795 0.8698

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
3
Safetensors
Model size
135M params
Tensor type
F32
·
Inference API
This model can be loaded on Inference API (serverless).

Finetuned from